Last October, I was petsitting for a friend and, while out for a walk with the pup, came across a strange looking bird (above). It was sitting, shell shocked, on the hood of a car in Brooklyn. When it was still there an hour later, I decided I would take it and release it in a nearby park—but how do you rescue a bird if you don't even know what kind it is?
No one on Twitter or Instagram could help; a friend who is a birder guessed that it might be a woodcock. So, after reading how to best rescue those birds, I gently wrapped the bird in a towel, slipped it into a paper bag, and walked it down to Brooklyn Bridge Park, where I let it out in some vegetation. It ran away from me, flying in short bursts, as fast as it could.
If only I'd had Birdsnap. This electronic field guide, created by computer scientists at Columbia University and the University of Maryland, can identify 500 common North American birds with nothing but a cell phone photo. The process is simple: Take a photo or choose one from your phone's album; click on the eye and the tail; and wait for the potential matches to show up.
After I'd set the bird loose in the park, my birder friend texted again: another birder believed the bird was a juvenile Virginia Rail, a freshwater marsh bird that mostly keeps to itself. Birdsnap identified the bird correctly on the first try. (How it got on the hood of a car in Brooklyn will forever remain a mystery.)
Columbia Computer Science Professor Peter Belhumeur and University of Maryland Computer Science Professor David Jacobs came up with the idea for the app when they realized that the software and techniques they'd developed for facial recognition could also be used to identify species. Facial recognition algorithms work by finding the resemblance between comparable parts of faces, comparing a nose to other noses and an eye to other eyes, according to Columbia's Engineering Department. In Birdsnap, each species has 17 parts marked; the app detects the parts of the bird so it can compare them with what's in its database and discover species that are visually similar to the animal in an uploaded photo.
"What's really exciting about Birdsnap is that not only does it do well at identifying species, but it can also identify which parts of the bird the algorithm uses to identify each species," said Thomas Berg, a Columbia Engineering computer science PhD candidate. "Birdsnap then automatically annotates images of the bird to show these distinctive parts—birders call them 'field marks'—so the user can learn what to look for.”
And the app, available for the iPhone, does more than just identify birds: It also provides descriptions of the animals and their calls, shows their family trees and similar species, and includes range and sightings maps.